[COPYRIGHT REGISTERED.] No. 138.
THE INSTITUTION OF POST OFFICE ELECTRICAL ENGINEERS
Telephone Cable Testing (including Fault Localisation).
BY
W. T. PALMER, B.Sc., Wh. Ex., A.M.I.E.E., and E. H. JOLLEY, A.M.I.E.E.
A PAPER Read before the London Centre on loth November, 1931, and at other dates at the following Local Centres:— South Western; South Midland; North Western and Eastern. No. 138
THE INSTITUTION OF POST OFFICE ELECTRICAL ENGINEERS.
Telephone Cable Testing (Including Fault Localisation)
By
W. T. PALMER, B.Sc., Wh. Ex., A.M.I.E.E., and E. H. JOLLEY, A.M.I.E.E.
A PAPER
Read before the London Centre on loth November, 1931,, and at other dates at the following Local Centres:— South Western; South Midland; North Western and Eastern. TELEPHONE CABLE TESTING (Including Fault Localisation)
SYNOPSIS. The paper is divided into three parts :— Part I. Deals with underground cables—tests required in the factory and during subsequent installation work and end-to-end or final tests—includes leakance measurements— laying and balancing tests—capacity unbalance measure- ment by direct method—groups of loading sections—non- repeatered and repeatered cables—near-end and distant-end cross-talk measurement—variation of cross-talk with fre- quency, length and type of circuit—impedance frequency and impedance unbalance tests—visual methods of measuring cross-talk, etc.—attenuation and impedance tests on music circuits. Part II. Classification of cable faults and localisation tests applicable—includes D.C. and A.C. methods—double- endAl tests—ballistic tests—method of mixtures—overlap— open and closed tests—slide-wire A.C. tests—Steven's test for C.R. faults—split pairs—split loading coils—obscure loading coil faults (short-circuited turns, etc.)—zero reactance test—cross-talk frequency methods. Part III. Deals with submarine cables—lead-sheathed and balata cables—factory tests—laying and final tests—fault localisation and repair operations—Mance, Kennelly tests, etc.
Note.—The small index numbers in brackets shown thus :—(1), (2), etc., are reference numbers which are listed at the end of the paper.
INTRODUCTION. With the continued increase in both importance and volume of international and inter-urban telephone traffic, the trunk and toll cable networks are playing. a more and more important part in the telephone system, and there is no need to emphasize the importance of the cable testing work which is necessary to secure (and subsequently to maintain) the high grade of transmission efficiency which is demanded. The 4 TELEPHONE CABLE TESTING. progressive improvements and refinements in telephone cable manufacture have necessitated corresponding developments of the testing operations, and it is part of the purpose of this paper to outline such developments. The cable testing methods used in the Department are surveyed and it is hoped that such a general review of the subject will not be without interest and value. Some of the more important test results are analysed and discussed, but space does not permit of every phase being investigated in detail, and references are generally given where a fuller description can be found. Certain cases are dealt with fully where technical research and development have enabled hitherto " standard " and/or relatively slow methods to be replaced by quicker and more satisfactory methods. The paper is subdivided as follows :— Part I. (A) Tests during manufacture. (B) Tests during installation. (C) Tests subsequent to installation. Part II. Fault Localisation Tests. Part III. Submarine Cable Tests.
PART I. (A).
TESTS DURING MANUFACTURE. The following non-electrical tests are carried out before completion of the factory lengths in the case of twin (ordinary, composite, distribution and aerial), multiple twin (P.C.M.T.) and star quad (P.C.S.0.) types of paper core cable. All the necessary conditions to ensure that the completed cable shall be satisfactory are reflected in the detailed requirements of the P.O. Engineering Department's Cable Specifications :—
(1) Conductors. These are visually examined for smoothness and all wires having a rough surface are rejected. By means of a micro- meter gauge the uniformity of diameter is examined at various points along the wire and if any appreciable difference is noted the wire is rejected. Joints are only permitted when a break occurs during the stranding process and must then be scarfed and soldered or welded for conductors over zo lb. gauge, and made in the presence of the Department's Inspect- ing Officer. TELEPHONE CABLE TESTING. 5 (2) Insulating Paper. The paper is measured for uniformity of thickness and is carefully examined for uniformity of texture and freedom from impurities. The breaking weight of the paper should exceed 4 lbs. for each inch width and o.00i inch thickness.
(3) Stranding. The colour scheme and the direction of the lays are examined to see if they agree with the relevant specification. When the length is completed the ends of the cable are opened and immersed in molten paraffin wax. The follow- ing further tests are then made :—
(4) Insulated Conductor. The insulating paper is tested for brittleness periodically by wrapping an unwaxed insulated conductor round a pencil. If satisfactory the paper should not split when subjected to this treatment.
(5) Lead Sheath. This is examined for mechanical defects and samples are periodically submitted for chemical analysis. The maximum diameter is measured with a micrometer caliper, care being taken to ensure that the diameter measured is the maximum in each case since the sheath may be oval. On a percentage of the lengths pressure tests are taken.
Electrical tests on the manufactured lengths are made to ascertain how far the pair and phantom circuits are satisfac- tory from a purely transmission point of view. These are :— Conductor resistance, dielectric resistance (both A.C. and D.C.) and mutual capacity. Measurements are also made of resistance unbalance, capacity unbalance, and in the case of continuously loaded cables, inductance unbalance, to deter- mine how far the circuits are satisfactory from the interference point of view.
(i) Insulation Resistance. The wires are grouped so that every wire is tested against all adjacent wires and the lead sheath. In the case of twin cables, the " A ' wires of alternate pairs in alternate layers are bunched to form a group, and similarly the " B " wires. 6 TELEPHONE CABLE TESTING.
The other pairs are similarly treated. This gives 8 main groups. Pairs which are not so included are separately grouped. Each group is then tested against all other groups and sheath. In the case of P.C.S.Q. and P.C.M.T. cables, the A, B, C, and D wires of the quads are first bunched together throughout the cable to give four groups and each group is tested against the other three and sheath. Then the four wires of alternate quads and alternate layers are grouped and tests made on each group against the remainder and sheath. The test is made with a galvanometer of the reflecting mirror type, the battery voltage is 300, and the galvanometer deflection is compared with that obtained when a standard megohm is substituted for the circuit under test.
(2) Mutual Electric Capacity (M.E.C.). This test can be taken when the cable is wired up in groups for the insulation resistance test, in the case of twin cables. The capacity can be measured between the A's and the B's bunched of each layer or between the whole of the A's in the cable bunched and the whole of the B's bunched. In the case of the composite twin cables the groups having different specified capacities are measured independently. P.C.S.Q. and P.C.M.T. cables (generally) have each pair circuit measured separately. The testing circuit is a simple form of Max. Wien bridge.
(3) Conductor Resistance (R). At least one pair in each layer and from 20 to 30 pairs in all (depending on the size of the cable) are connected in series for the test. In the case of composite cables at least to pairs of each size of conductor are measured. The test is made with a 4-dial Wheatstone Bridge. A correction to R must be made (i) for the lead resist- ance, (ii) for the temperature.
(4) D.C. Resistance Unbalance. This test is one for readily obtaining the difference in resistance between the A and B wires of a pair, expressed as a percentage of the total loop resistance. The testing circuit is arranged so that the unbalance can be read directly from a graduated scale as a percentage of the TELEPHONE CABLE TESTING. 7 loop resistance.0) The tests are only carried out in the case of main underground cables.
(5) Capacity Unbalance. This test is made in the case of P.C.M.T. and P.C.S.Q. cables to determine the inequalities of wire-to-wire and wire- to-earth capacities. These inequalities (which form the principal cause of cross-talk) are usually expressed in micro- micro-farads. The testing circuit is generally some form of A.C. capacity bridge. The measurement of capacity unbalance is considered later in connection with installation tests.
(0) Leakance. Leakance is the reciprocal of the effective resistance of the dielectric and is denoted by G. The ratio G/C, i.e., Leakance Capacity, is referred to as the Leakance Con- stant. As the power factor for paper core cable dielectric is so small the ratio G/wC gives the value of the power factor nearly enough for all practical purposes. At Soo p.p.s. the power factor of the air spaced paper core dielectric of a typical underground cable is about .003 (G/C = 15), for a solid paper core submarine cable it is about .005 (G/C = 25), while for submarine cable having gutta percha or balata as dielectric it varies from about .02 to .0 (G /C = too to 5o). Of recent years submarine synthetic dielectrics have been manufactured with the ratio G/C as low as 7, for example, paragutta, a dielectric which has been proposed for the Trans- Atlantic Telephone Cable. Despite such low values of leakance, however, it is important that its value should be accurately determined during the process of manufacture and this has been made more necessary since the adoption of ink line markings on the paper for identification purposes. It has been found that unless the ink has been carefully chosen it can introduce con- siderable losses in the dielectric. On account of the low value of the leakance in relation to the other primary con- stants, its accurate measurement is attended by some difficulty. In the first place, to secure accuracy and to avoid the necessity of corrections on account of the resistance of the wires, the measurement must be made on short lengths of cable, e.g., in 8 TELEPHONE CABLE TESTING. the case of underground telephone cables the measurements are made on lengths of about 200 yards and, in the case of submarine cables, on lengths of about 6o feet. Measure- ments at one frequency only need be made in a routine test, usually at Soo p.p.s., on a small percentage of factory lengths of each cable. There are several methods of carrying out the necessary tests and they all include elaborate attempts to overcome the difficulties particular to the problem. The testing set described below has been used in recent experimental work on submarine cable cores. It is a set which avoids the use of very expensive high-grade condensers having extremely low power factors which are necessary in certain methods. It pro- vides an accurate and rapid means of determining the leak- ance constant and is therefore particularly suitable for routine factory tests. Since it is usual to refer to the ratio G/C in dealing with dielectric properties, the set is referred to as a " G/C Bridge." G/C Bridge. Fig. I shows the principle of the method of measurement. The unknown admittance is connected to points B and C.
FIG. I.
P, P, and Q, Q, are non-reactive fixed resistances. K is an adjustable condenser. R is a non-reactive adjustable resistance. Balance is secured by the adjustment of K and R and the conditions for balance are as follows :— C = K (I) TELEPHONE CABLE TESTING. 9 R G (2) Q(Q + R) It will be seen that, if R is very small compared with Q, (2) may be written :— R G =
In practice, Q is made equal to io,000 ohms (and, so long as R is not greater than about ioo ohms, the foregoing approximation is sufficiently accurate for practical purposes) and hence G in micromhos is given by loo The simple form of bridge shown in Fig. I is, however, not suitable for the accurate measurement of the low leakance values met with in telephone cables. In the first place it necessitates an accurate calibration of the power factor of the standard condenser, (K), or, alternatively, the use of a very costly high-grade condenser, such as an air condenser using silica-quartz mountings, of which the power factor may be neglected. Errors are also introduced (I) by inequalities in the values and distribution of the capacity couplings and leakance paths between the components themselves and earth, and (2) by inequalities in the reactive components of the resistances, which would be quite negligible under general A.C. testing conditions. Errors due to (I) can be eliminated by an elaborate system of screening as shown in Fig. 2, which also enables the equalising to earth of the points B and C to which the unknown admittance is connected.(2) Referring to Fig. 2 it will be seen that all the components, including the secondary windings of the transformers, in the supply and detector circuits, have an inner screen which is connected to one end of the component. This ensures that all capacity and leakance paths from the component terminate at some definite point. In the case of the ',ow ohm ratio- arms this is point A of the bridge. For the io,000 ohm and adjustable resistances, the secondary of the input transformer and the two condensers KJ and K2, this point is B. The inner screen on the detector transformer secondary is con- nected to C and is continued from the transformer to point D of the bridge and its associated connections. This ensures I0 TELEPHONE CABLE TESTING.
Detector
Double Screen Transform ----- . -3211.0.012.4------
1. ,...••••=111 0 Earthed Screen Unknown Admittance.
FIG. 2. that point D shall only have capacity or leakance to point C and, as this will be in parallel with the detector circuit, it will not affect the balance of the bridge. Similarly, the inner screen on the input transformer secondary is continued from the transformer to point A of the bridge and its associated connections, which in this case include the inner screens of the ',ow ohm resistances. This ensures that point A of the bridge shall only have capacity and leakance to point B and as this will be in parallel with the source it will not affect the balance of the bridge. The residual capacities and leakances of the bridge are by this means located to points B and C and it only remains to fix these values, subsequently taking account of them by means of an initial balance of the bridge. TELEPHONE CABLE TESTING. II
The method adopted is to enclose the components of the bridge in earthed screens, or to enclose the whole of the bridge in a single earthed metal screening box. It will be seen that the use of double-screened transformers is involved and these prevent any unbalance to earth of the source and detector circuits from affecting the balance of the bridge. With the completion of the screening of the bridge, points B and C only have admittance to one another and to earth. The method of connecting up the screens, however, ensures that the capacity of B to earth is greater than that of C to earth and it is therefore a simple matter to increase the capacity of C to earth by the addition of a condenser, as shown, and thus make it equal to that of B. This equality is a matter of importance when the bridge is used for measure- ments of admittances, such as that of a cable pair, which are essentially- balanced to earth and should remain so during the measurement. Errors due to (2), i.e., differences between the reactive components of the ',pop-ohm and io,poo-ohm resistances, are overcome by a method due to Dr. L. G. Brazier, which also avoids the necessity for the use of a standard condenser of known or negligible power factor.(3) According to this method, a fixed condenser, having a capacity somewhat greater than the maximum value of capacity which it is re- quired to measure, is connected across the arm BD of the bridge, and a variable condenser, having a maximum capacity equal to that of the fixed condenser, is connected across the arm BC. When the admittance to be measured is connected to the points B and C, the bridge is balanced by reducing the capacity of the variable condenser by an amount equal to the unknown capacity, and increasing the value of the resistance by an amount R until the total admittance between the points B and C is the same as before. The bridge solution already given holds also in this case, viz. :— R = G (in micromhos). 'Po The total admittance of each of the bridge arms remains unaltered and the method is thus essentially one of substitu- tion and errors due to the aforementioned inequalities are eliminated. The leakance of the condensers used does not require to be known, since this is taken into account in the initial balance of the bridge. All that is of importance in this con- 12 TELEPHONE CABLE TESTING. nection is that the leakance of the variable condenser should not vary with change of setting of the condenser. This con- dition is sufficiently fulfilled by the use of well designed continuously variable air condensers in which the only losses of importance are those in the solid dielectric used in the mounting of the plates, which losses in recently constructed condensers are essentially constant for all settings of the condenser. With the bridge used in the experiments, the change of leakance of the variable condenser, for a given change of capacity, is less than the leakance of a silica-quartz air condenser of corresponding capacity, which is itself less than can be measured by any known means of calibrating such condensers and is therefore negligible for all practical purposes. A G/C bridge was constructed on the above plan for the Department by Messrs. Gambrell. Constancy of calibration of the bridge with reasonable constancy of temperature has been secured by placing the whole of the components in a well constructed case and Fig. 3 is an external view of the complete instrument.
EXTERNAL VIEW OF G/C BRIDGE. FIG. 3. "
Admittances having a capacity up to 14,000 µµF and leakance up to 11.i micromhos can be measured and, by using TELEPHONE CABLE TESTING. 13 an amplifier and telephone receiver in the detector circuit, a sensitivity of o.00t micromhos can be secured with an accuracy of 1% which is well within the limit demanded by practical considerations. The bridge can be used over a range of frequencies from 30o to 9,000 p.p.s. For tests above 3,000 p.p.s., in place of a telephone receiver an amplifier-rectifier and a sensitive D.C. galvanometer have been used in the bridge detector circuit.
PART I. (B).
TESTS DURING INSTALLATION. (t) Laying and Balancing Tests. Telephone cables are subjected to electrical tests during all stages of the laying operations. In their simplest form these tests are merely : — (a) For continuity and freedom from earth or contact. (b) To prove absence of crosses. (c) To ascertain insulation resistance. (d) To ascertain conductor resistance. (e) To prove absence of overhearing. Tests (a) and (b) are made with a battery and lineman's detector ; (c) with a megger ; (d) with a Wheatstone bridge ; and (e) with a buzzer or telephone as the source of disturbance, disturbing on individual or bunched pairs and listening on other pairs with an ordinary telephone. In the case of toll and the more important loaded junction cables these tests are supplemented by capacity unbalance and cross-talk measurements on loading sections. Recent ex- perimental work has been carried out in connection with aerial cable, using to lb. conductors, systematically jointed* (i.e., no capacity balancing in the field), coil loaded, and worked " four-wire " with the object of using such repeatered cables for toll circuits instead of 4o lb. or 7o lb. non-repeatered loaded underground cables, as at present used. With aerial cables worked in this manner the amount of testing required during installation is considerably reduced, consisting
. Systematic jointing is a method by which pairs or quads are crossed at the joints in accordance with a predetermined sequence to reduce as far as possible the length over which any two circuits will be adjacent through- out the cable and was introduced in 1926 for use in loaded twin cables. 14 TELEPHONE CABLE TESTING. essentially of insulation and conductor resistance and cross- talk tests on completion.(') The tests imposed during the laying of main trunk under- ground cables are as follows :— (a) Insulation Resistance. (b) Conductor Resistance and Conductor Resistance Unbalance. (c) Capacity Unbalance. (d) Cross-talk (terminated). (e) Mutual Electric Capacity. Test (a) is made with a 50o volt megger. Test (b) is made using a specially designed Trunk Cable Resistance Test Set(') which, besides giving a more accurate measurement of loop resistance than the ordinary P.O. Wheatstone Bridge, provides for the direct measurement of the conductor resist- ance percentage unbalance. Test (d) is made by using the standard P.O. cross-talk testing apparatus(1) which includes a reed-hummer as the source of disturbance, and a Western Electric cross-talk meter for measurements.(5) The cross-talk tests made in the installation stage on loading sections of balanced cables are generally to replace, for the sake of speed, certain capacity unbalance tests for determining the interference between circuits in different quads, when accurate determination of the capacity unbalances is not required. Test (c) (Capacity Unbalance). The method of measur- ing capacity unbalance by means of the " double bridge,"19 (which involves calculations from the bridge readings for the determination of the required capacity unbalance character- istics) has now been largely superseded by a method in which the required interference characteristics are given directly by the bridge readings, thus speeding up the balancing opera- tions. The principal characteristics involved are :— Phantom to Side (producing phantom to side cross- talk). Side to Side (producing side to side cross-talk). Side to Earth (producing earth interference). These refer to circuits in the same quad (" within quad "). Other characteristics refer to the interference between circuits in different quads (" between quads "). Taking any two quads in a cable there are 15 different capacity unbalance characteristics (if the phantom circuits are TELEPHONE CABLE TESTING. 15 taken into account) and if the phantoms are excluded (as when phantom working is not required) the number is reduced to 7. With the direct method, a switch, due to Mr. H. T. Werren, having 15 positions, is used in conjunction with the testing bridge which enables a further considerable speeding up of the testing work, especially when the " between quad " characteristics have to be measured. The following is a brief outline of the method(6) :— Fig. 4 shows the components of the bridge which consist of two 1 ,000 ohm non-reactive ratio arms, two 600 µIII' fixed air condensers and two variable air condensers 0 —1200 ,upF.
RING SUPPLY EARTH 00 00 0 0 A. -0 -0 0 B. 0 C. 0 DI <1 00 0 A2 0 82 0 Cl 0 DI =0=. 00 — TEL.
RING SUPPLY EARTH 00 00 0
FIG. 4.
The two latter condensers read zero when set to 600 ,up,F, that is, when they balance with the fixed condensers. The balance of the bridge is disturbed when unbalanced circuits are con- nected to it and balance is restored by increasing or decreas- i6 TELEPHONE CABLE TESTING. ing the capacity of the variable air condensers. The dotted connections in the lower figure are used for initially balancing the bridge when the scales of the condensers are adjusted to allow for slight inaccuracies in the bridge components. In Fig. 5(a) the capacity network of a cable quad is shown and the usual nomenclature of direct capacities w, x, y, z, a, b, c, d, m, it, is used, while Fig. 5(b) shows the net- work reduced to an equivalent 6-branch network by applying the " Network " or " Star-Mesh " transformation theorem.(") When the capacity network is considered in this form the measurement of the various capacity unbalance characteristics by the direct means can be easily followed.
vi—x=p z — y=q (a) ral — Z = r x—y=s a —b=u c — d=v
(b)
B CAPACITY NETWORK. FIG. 5.
Fig. 6 shows how the wires are connected to the bridge for the measurement of the phantom to side, side to side and side to earth characteristics and the resulting disposition of the cable capacities. Considering the phantom to side case it will be seen that the source is across one side circuit (AB) and the telephone is, in effect, connected across the phantom circuit. Then the condition for silence in the telephone will be seen to be :— (a — b)(c + d) k, — Ki = (w — x) + (7, — Y) ± a+b+c+d TELEPHONE CABLE TESTING. 17
Phantom Side,
Side Side
Side Earth
FIG. 6. which may be written k i — K1 =p+q+?1U -where p, q, u have the values shown in Fig. 5 and a ---->- b — - > - - c d The reading on condenser K„ therefore, gives the value p + q + u
18 TELEPHONE CABLE TESTING.
which is known as the phantom to side interference character- istic. Condenser K2 is used for what may be termed a power factor adjustment. Similarly, the side to side interference characteristic can be measured directly by K,, as shown in Fig. 6, K, being first set to read zero so as to balance k1. This gives
(a-- b)(c — d) k, — K, = (w — x) — (z y) a+b+c+ d i.e. , k, — K, p - q. (a — b)(c — +d neglecting The reading on the condenser a + b + c d . K1, therefore, gives the value P - q which is known as the side to side interference characteristic. Referring to the side to earth case it will be seen that a considerable capacity is thrown in parallel with the ratio arms, namely, w + z in one case and x + y in the other. The difference of these two expressions is p + q. When testing on jointed lengths of cable, if the circuit under test has been balanced for phantom working, p + q will not be large; and the presence of the capacities mentioned, in parallel with the ratio arms, will not as a rule lead to any difficulty. If, how- ever, the cable has not been specially balanced for phantom working, p + q may be fairly large, and some difficulty may be experienced in obtaining a satisfactory balance, especially on the longer lengths of cable. In addition, the irregular dis- tribution of the (p + q) unbalances will give errors in the value obtained for the characteristic, for which it is not possible to allow. In the case of lengths of cable of more than 2,000 yards it has hitherto generally been the practice, (when the cable has not been balanced for phantom working) to rearrange the components of the bridge and measure the side to earth characteristics in the standard double bridge method, but another method by which the error can be avoided has been suggested by Mr. Hodge of the Research Section. Fig. 7 refers. An additional switch S is included in the testing circuit as shown in Fig. 7(a). With this switch at position i the testing conditions are normal. In order to measure the un- balance to earth of the AB side circuit (i.e., " u "), the TELEPHONE CABLE TESTING. IO
Bridge
h
/12 ic Sw/ l 4 5 2 2 02 os p 5 / on ls ina
(a) m 5 r